Well i need to learn a bit of C soon.
After my server is running Linux now, could it come very useful.

"J.O. Aho" <user@example.net> skrev i en meddelelse
news:3ao9blF5gn2k7U2@individual.net...
> Denni Pultz Gottfredsen wrote:
>> Had looked into coding..
>> Im coding html, php
>
> Then you are always there, php isn't that far from C IMHO, but you coulr
> write a PHP script that works as identd and use xinetd to power it.
>
> Or just take it easy, sit infront of telly and know that fakeidentd will
> work as it is.
>
>
> //Aho

On Sun, 27 Mar 2005 15:30:35 +0200,
"J.O. Aho" <user@example.net> wrote:

> As I said before, it don't have much of options, if you want to change
> username that it will respond, then add it to the line where you start
> fakeidentd

Which opens up another option. Can you get your client to (re-)start
fakeidentd each time it starts a new connection? Or maybe add a small
command to the client to do it manually. "/fake mynewid", or something.
Most decent IRC clients have half-way decent scripting capabilities.
X-Chat that I use can script in TCL (which I know from building up
bots), Perl, and something else I think... Just poke around the docs,
and it shouldn't be that hard to put something together. Then set up
your sudo config to let you run fakeidentd without a password (it's not
a good idea to disable password totally in sudo, in case someone finds a
way to abuse your script to run their own commands -- which I've seen
happen before), and use that to kill and restart fakeidentd with
whatever name you want. (I'd kill it again once the MOTD is finished,
personally. Probably even earlier.)

I used to do something like that years ago at uni (damn... was it
really THAT long ago?), before I knew about oident and fakeident and so
forth. A friend of mine was running a machine on which he and several
others in his department were working on a little project of theirs, and
they relied on identd giving the proper information. But me and the guy
running the thing wanted to be able to fake our IRC idents (without
interfering with our participation the project, of course). Now, the
stock identd had the ability to check for a ".noident" file in the users
home directory, and reply with a standard "this user doesn't want to be
idented" message. So I patched a little extra in right after the check
so it would also search the file for a line describing that particular
connection (with some very basic wildcard support), and use
username/number provided by the last matching line (with anything else
on the line ignored so we could drop in markers). So we just added a
little scripting to our clients that read the current file, an re-wrote
it with a catch-all up top that gave the proper response, followed by a
line just for the current connection with the fake id's.

I'm not saying you have to go to quite that extent (I wish there was an
ident server that did, though, because I can't be bothered re-hacking my
current one ), but the basic idea of running fakeidentd for a short
period (since the ident scan usually isn't repeated once the connection
is established) from the client itself should do the trick, as long as
only one person will be connecting to IRC at a time.


Fredderic